Based on records from the Federal Bureau for Statistics of Germany, the sea
sonality of mortality was investigated for the period 1946-1995. Lowest mor
tality rates were found during summer (August or September) while highest v
alues were found in winter (January through March). Non-linear regression o
f all monthly mortality data with the average monthly temperatures in Germa
ny revealed a significant negative relationship (r=-0.739; n=600; P<0.0001)
. The fact that the differences between the long-range monthly temperatures
and the individual monthly temperatures also showed a distinct relationshi
p to the mortality rates speaks against a mere coincidence of both paramete
rs. The amplitude of this seasonal rhythm declined steadily within the obse
rvation period. It is concluded that low temperatures cause an increase in
mortality rates and that this effect has become less important during recen
t decades due to the increased use of central heating and because of improv
ements in the public health system.